As men age, both their bone mineral density and their serum testosterone concentrations decrease, and increasing evidence suggest that the decrease in testosterone might be a cause of the decrease in bone mineral density. We therefore propose to test the hypothesis that increasing the serum testosterone concentrations of healthy elderly men to those of healthy young men will improve their bone mineral density. We plan to recruit healthy men >65 years from the life care communities and senior centers associated with the Philadelphia Gerontology Research Consortium and to screen them to include those who have a serum testosterone concentration and a bone mineral density < mean - 1 SD of men 20-40 years old and exclude those who have identifiable bone diseases or diseases that would preclude compliance with the study protocol. We estimate that we shall need to screen 300 volunteers to find 100 to begin the protocol and have 70 complete it. The estimation that 70 subjects will need to complete the protocol to obtain meaningful results is based on several assumptions, including the precision of the procedure, and on the judgment that an improvement of greater than or equal to 4% in bone mineral density would be clinically important. The study will be double-blinded, and subjects will be randomized to either a testosterone treatment or a placebo group. Testosterone will be administered transdermally, in order to produce a physiologic pattern of serum testosterone concentration and to facilitate recruitment of subjects. The serum testosterone and SHBG concentrations throughout the study will be measured, and signs of adverse effects of testosterone (prostate cancer, BPH, sleep apnea) will be monitored. The principal index of the effect of testosterone will be bone mineral density of the femoral neck and trochanter, measured by dual energy x-ray absorptiometry (DEXA). Any difference in bone mineral density between the testosterone and placebo groups at the end of the study will be interpreted in light of measurements of indices of bone metabolism, including urine deoxypyridinoline and hydroxyproline and serum osteocalcin and PTH. We expect that the results of this study will help determine if increasing the serum testosterone concentrations of healthy elderly men to those of healthy young men will improve their bone mineral density and, if so, at what risk of increasing the incidence of testosterone-dependent diseases.